We would need a scope shot to confirm that. It the amp is oscillating at >> 20 kHz you’d probably not hear it.
The Zobel resistor is not heating. The resistor that is vibrating and making the sound is the pair of 0.1R BPR 5W resistors.
At first I heard the noise near the Zobel resistor and capacitor. Later by touching the Zobel resistor and capacitor and later the BPR series resistors right next to the Zobel I found out it was not the Zobel but instead the BPR series resistors.
Again it only occurs during test and there is no noise or output signal when the generator is off. I suspect that if the ceramic cased BPR resistor was replaced with a metal film then it would be silent. The 0.003% THD measurement does not show evidence of oscillation.
At first I heard the noise near the Zobel resistor and capacitor. Later by touching the Zobel resistor and capacitor and later the BPR series resistors right next to the Zobel I found out it was not the Zobel but instead the BPR series resistors.
Again it only occurs during test and there is no noise or output signal when the generator is off. I suspect that if the ceramic cased BPR resistor was replaced with a metal film then it would be silent. The 0.003% THD measurement does not show evidence of oscillation.
That is your first clue. The transistors were fakes. It stands to reason that the 0.1 ohm resistors that are making racket might be fake (or just crap quality) too. If the winding is loose enough to rattle from the current produced by 64 watts of output they will be at risk of eventual metal fatigue. Put new resistors in there and call it good.
Yes. I will replace those. I have replaced the outputs with genuine devices. Next I wonder about the drivers. Originally the L20.5 shipped with a better driver part number. That has been silently downgraded and I also don't know if the installed drivers are fake or not. That might also be part of the THD being higher than the seller quotes.
Doesnt this use 2SB817/2SD1047 ? The 817 is out of production now.. so yeah... NJW3281/NJW1302 would work well i guess
I consider the 2SB817/2SD1047 in these kits highly suspect. The marking does not look like any genuine manufacturer. Also they blow up regularly. Both the L20 V7 and the L20.5. With eight output devices that is ridiculous. With genuine Toshiba they have never blown up. I even took one blown up L20.5 (that is four pairs of outputs) and "test" repaired it with one pair of genuine Toshiba 2SC5200N/2SA1943N from Mouser. Did not blow up. Later I add the remaining pairs.
In another L20.5 thread there are L20.5 buyers who have gone through board after board.
Today I will work on the other board and I might use the NJW0281/NJW0302 or the MJL3281/MJL1302. The MJL3281/MJL1302 are physically hard to fit in due to the heatsink and PCB configuration so I will likely use the 0281/0302.
Next I might try genuine Toshiba 2SC5171 & 2SA1930 drivers salvaged out of one of the old Onkyo home theater receivers that I use for a chassis and power transformer.
I don't expect to see much of a difference but I am curious to see if I can detect a THD different from the different outputs and drivers. Perhaps not but I enjoy these experiments.
In another L20.5 thread there are L20.5 buyers who have gone through board after board.
Today I will work on the other board and I might use the NJW0281/NJW0302 or the MJL3281/MJL1302. The MJL3281/MJL1302 are physically hard to fit in due to the heatsink and PCB configuration so I will likely use the 0281/0302.
Next I might try genuine Toshiba 2SC5171 & 2SA1930 drivers salvaged out of one of the old Onkyo home theater receivers that I use for a chassis and power transformer.
I don't expect to see much of a difference but I am curious to see if I can detect a THD different from the different outputs and drivers. Perhaps not but I enjoy these experiments.
LJM schematics are a curious affair. They often contain deliberate mistakes (I am told) to foil copycats. I never did find a real one for the L20 V7 (rare CFP version) or the L20.5. The ones I found were clearly very wrong. The seller provided a non-CFP output type schematic for the L20 V7, for example.
I have been using the L12-2 schematic.
Does anyone have a checked/accurate L20.5 schematic that they can post?
Original thread link: L20.5 AP SYS Test data My design
Believe it or not
Believe it or not, but LJM has "lost" schematics of all his "creations" but he never lost a single Audio Precision chart indicating very special, almost unbelieveable performance. A great magician with all these mirrors and smoke. A master. 😀
Believe it or not, but LJM has "lost" schematics of all his "creations" but he never lost a single Audio Precision chart indicating very special, almost unbelieveable performance. A great magician with all these mirrors and smoke. A master. 😀
Last edited:
Well YJ seems to use better quality components in his products. However the performance is not as good and the PCB traces lift. (L28 as an example.) LJM seems to have better performance (I mean what I have measured) and the PCB quality is good. But the components (outputs) blow up. So YJ has better transistor brands/quality. LJM has better capacitor brands, PCB and performance. (However I really wonder about some of the capacitors on the LJM capacitance multiplier I bought... That is another story.)
Sigh.
Don't laugh too hard but the Toshiba 2SC5200/2SA1943 that I used to repair the LJM L20.5 yesterday came from the YJ L28.
Sigh.
Don't laugh too hard but the Toshiba 2SC5200/2SA1943 that I used to repair the LJM L20.5 yesterday came from the YJ L28.
Last edited:
I couldn't resist, YJ probably gets his component from ICS scrapyard.
I usually purchase just bare boards or kits and throw away all components. Right now I have ALL ON-Semi transistors from Farnell for my favourite clone board.
My next project will be AKSA & XRK's Alpha Nirvana, to avoid all that mess. 😎
Last edited:
When I decide on my favorite and buy a nice chassis (and nice Toroidal transformer too) I will likely just use the PCB and all new components.
But for the fun of exploring the sound and measurements of different types (such as QUAD405-2 vs L20.5 CFP vs MX50X2) the LJM kits are accessible. Probably fine at +/-40V but not higher in my experience.
But for the fun of exploring the sound and measurements of different types (such as QUAD405-2 vs L20.5 CFP vs MX50X2) the LJM kits are accessible. Probably fine at +/-40V but not higher in my experience.
Dear kozard,
I gree with you: I have learned a lot at low cost (damage) from building these cheap amplifiers. I've analysed them quite extensively, learned how to troubleshoot problems and how to find adequate substitutes. To "resurect" dead LJM boards I've used various ISC substitutes which worked rather well for that purpose. I am preparing now a curve tracer adapter for my oscilloscope with aim to analyse and pair transistors. In the end I will make my final triamping system. I've always been determined to achieve my goals. This time too.
Thanks for your inspiring observations.
Cheers my friend🙂
I gree with you: I have learned a lot at low cost (damage) from building these cheap amplifiers. I've analysed them quite extensively, learned how to troubleshoot problems and how to find adequate substitutes. To "resurect" dead LJM boards I've used various ISC substitutes which worked rather well for that purpose. I am preparing now a curve tracer adapter for my oscilloscope with aim to analyse and pair transistors. In the end I will make my final triamping system. I've always been determined to achieve my goals. This time too.
Thanks for your inspiring observations.
Cheers my friend🙂
As power transistor fakes seem to be the most common problem for DIYs and LJM has been aware and freely discussed this here, over many years now, you should be aware that many cheap LJM kits and even kits with his trademark have been fakes too, along with the fake semis too, of course.
'First thing to notice is that LJM's own store is at Tao Bao, the Alibaba Chinese language platform intended for local shoppers. So we can buy his product from other sellers at other platforms too but not easily from LJM directly so instead, we typically buy the cheaper but faked kits. Actually, I have found and bought a few genuine, LJM pre-assembled kits, most recently the SMT Preamp 9 but the thru-hole parts kits sold at Aliexpress and Ebay have turned out to be full of fakes or just sub-standard components.
It seems it's price that drives the fakes market and we are the suckers who fuel it. If you want the genuine recommended power semis, KEC still manufacture and supply good copies of D1047 and B817 at low prices but as other posts suggest, you'll probably find it easier to just go to regular distributors like Digikey, Farnell, Arrow, Reichelt etc. websites and pay their top dollar prices for genuine On-semi equivalents.
'First thing to notice is that LJM's own store is at Tao Bao, the Alibaba Chinese language platform intended for local shoppers. So we can buy his product from other sellers at other platforms too but not easily from LJM directly so instead, we typically buy the cheaper but faked kits. Actually, I have found and bought a few genuine, LJM pre-assembled kits, most recently the SMT Preamp 9 but the thru-hole parts kits sold at Aliexpress and Ebay have turned out to be full of fakes or just sub-standard components.
It seems it's price that drives the fakes market and we are the suckers who fuel it. If you want the genuine recommended power semis, KEC still manufacture and supply good copies of D1047 and B817 at low prices but as other posts suggest, you'll probably find it easier to just go to regular distributors like Digikey, Farnell, Arrow, Reichelt etc. websites and pay their top dollar prices for genuine On-semi equivalents.
You're right Ian, as usual,
A Hi-End Audio system at price below 50 bucks is an offer one cannot refute. Irresistible.
They understand us perfectly. It is up to every individual to understand and refute such a "bargain".
The vibration of the 0.1 Ohm BPR 5W resistors is the same frequency as the test signal. Yes it does track the input test signal.
There is no output signal and no whistle when the input generator is off.
At this point I just plan to buy and install better resistors.
Due to the design of this four output pair CFP amplifier there is only one pair of 0.1 Ohm BPR 5W resistors. (And eight 50mOhm resistors, one for each output transistor.)
Any suggestions on a better replacements for such a high power amplifier?
There is no output signal and no whistle when the input generator is off.
At this point I just plan to buy and install better resistors.
Due to the design of this four output pair CFP amplifier there is only one pair of 0.1 Ohm BPR 5W resistors. (And eight 50mOhm resistors, one for each output transistor.)
Any suggestions on a better replacements for such a high power amplifier?